Tri-Metric program question [RE-wrenches]

[Allan Sindelar]

Wrenches,
I had promised to post the results of some Tri-Metric "tweaks" we were
testing out. On two jobs, the following has proven to work well for several
weeks now. See below for the original issue; short form is how to trick a
Tri-Metric amp-hour meter into holding its percent value while the grid is
up, yet being reasonably accurate during an outage.

The main trick was to set the battery efficiency to 100%. It means that the
Tri-Met will measure any outage discharge accurately, but will be a few
percent optimistic during recharge. In most cases, and on most systems, this
is OK for the occasional short-term outages most GTWB systems are designed
for. For longer outages, there's only an accumulated error issue when there
are multiple charge/discharge cycles without meeting the charged parameters.

We also set the voltage to just above float (52.4V with Concordes) and
charged amps to one A, but we are not yet sure how significant these
settings are to the success of the tweak. Note that we only changed Tri-Met
settings, not system settings, as we wanted to try a solution that we could
set up over the phone with the homeowner.

So far, both test systems are maintaining 100% indication during normal
(grid up) operation, yet count down normally during a simulated outage.

Allan at Positive Energy

----- Original Message ----- 
From: "Allan Sindelar" <allan at positiveenergysolar.com>
Sent: Wednesday, October 18, 2006 4:59 PM
>
> Wrenches,
> I posted this question a week or two ago, and was surprised that a) others
> had faced the same issue, b) nobody had worked out a simple, effective,
> proven solution. While I respect Mick's suggestion of a relay-based B+
> control on the monitor as probably working just as he envisioned, the
extra
> cost and complexity make it not the first approach to try.
>
> So the answer appears to be somewhere in a combination of these four
> approaches:
>
> 1) Set the float voltage slightly high, in order to increase charge
current
> relative to discharge current during float/sell conditions. This will
lessen
> accumulated losses that cause the meter to slowly lose %-of-full
indication
> (thanks for this idea, Todd).
>
> 2) Set the charge efficiency factor at or close to 100%, so that routine
> charge-discharge cycling has less of a tendency to accumulate as an
> inaccurately low %-of-full, or takes longer to do so.
>
> 3) With an Outback system that has a Mate, the AC IN hot button will
easily
> access a "force bulk or float" command, which initiates an automatic bulk
> charge cycle. I can instruct the homeowner to do this occasionally - say,
> every couple of weeks, or when the monitor is reading too far out of
whack.
> This is an especially good step for a system with flooded batteries, as
they
> otherwise only vary between float/sell and resting voltages, and a
periodic
> bulk charge cycle would be good for them. It's probably only useful with
> flooded batteries and involved homeowners (who are already caring for
> flooded batteries).
>
> 4) It might work with sealed batteries to set the Tri-Metric charged
voltage
> parameter below the float/sell voltage, and the charged current parameter
> just above the current necessary to maintain float. That way, the Tri-Met
is
> held at 100% and is constantly being reset as long as the grid is up.
> Following or during an outage, the low charged current parameter keeps the
> meter from resetting until charging current drops way down, even though
the
> voltage setting is artificially low. I wonder how these settings would
> affect monitor accuracy during an outage - I suspect not much.
>
>
>
> Feedback? Preferences among these options? What combinations would you
try?
>
>
>
> Thank you,
>
> Allan at Positive Energy
>
>
>
> Excerpts from previous posts about this issue:
>
>
>
> We have been including a Tri-Metric monitor with our Outback
> grid-tie-with-battery-backup systems, so that the homeowner has some sort
of
> state-of-charge indication during an outage. We have learned to be
selective
> about whether to include a Mate as well, as it's not as user-friendly for
> our typical non-technical grid-tie customer.
>
> The issue is that the Tri-Met is fundamentally designed for off-grid use,
> and uses charged-voltage and charged-current parameters to reset the
monitor
> on a regular basis. As the PS1 keeps the batteries in float, the Tri-Met
can
> develop an accumulated error that is only reset after a grid outage and
> recharge cycle, which could be months at a time. We have seen this happen,
> where the % of full slowly drifts down over days or weeks.
>
> I called Ralph Heise about this, and he had suggestions, but no clear
> answers - in fact he is looking for ideas that have worked too. He
suggested
> the following two ideas, which make sense in theory:
> 1. Set the charged voltage parameter below the float voltage (far enough
> below to accommodate the effect of hot-weather temperature compensation -
> say 52.4V if float is at 53.6 - and set the charged current setpoint just
> above the float current - 1 or 2A, I would guess. This keeps the monitor
> reading 100% while in float, and the combination of voltage and current
> settings would prevent premature resetting based on voltage-above-float
> alone.
> 2. Set the charge efficiency factor unrealistically high - say at 100% -
so
> that this keeps the % reading at 100% while in float. This would make the
%
> reading slightly high during outage cycling, but it could be easily reset
> each charge cycle.
>
> Has any Wrench out there come up with a good solution? How would you set
up
> a TriMet for this application? Or otherwise, how do you address the bigger
> issue of a monitor that is used only during an outage?
>
> ________________________
>
>
>
> If you are having problems with the amp hour meter on a grid, tied,
> battery based system, try raising the float voltage slightly on the
> inverter/charge controller. Yes, it will slightly lower the efficiency
> and cause the batteries to consume a bit more water, but the slightly
> additional charge current will offset the discharge pulses and keep the
> meter on track... also from what I have read, slightly overcharging wet
> cell lead acid batteries does more good than harm. I have also found
> that a few charge/discharge cycles will get the meter to recalculate
> battery efficiency higher (the e-meter starts out a default of 90%) so
> the +/- pulses are treated closer to being equal.
>
> When a battery based grid tie system has the grid operating, the meter
> is not necessary. When the grid goes down is when the customer needs the
> meter and it is then that it is important for the meter to read
> accurately, starting with a 100% charged reading.
>
> Grid tie inverters with batteries constantly send small charge and
> discharge currents into the batteries. Because batteries are not 100%
> efficient, amp hour meters treat charge currents and discharge currents
> differently. It takes more charge current to compensate for a given
> amount of discharge current. This is why amp hour meters on battery
> based systems tend to accrue a discharged SOC inaccuracy.
>
> The only way I have found to resolve this issue, so when the grid goes
> down the meter starts at 100% is by setting the float/sell voltage
> slightly higher so the batteries constantly get way more charge then
> discharge currents. Another thing which helps is to cycle the batteries
> a few times so the amp hour (link-10) meter recalculates the battery
> efficiency higher from the default of 90%.
> Todd
>


- - - -
To send a message: RE-wrenches at topica.com

Archive of previous messages: http://lists.topica.com/lists/RE-wrenches/read

List rules & how to change your email address: www.mrsharkey.com/wrenches/etiquette.php

Check out participant bios: www.mrsharkey.com/wrenches/

Hosted by Home Power magazine

Moderator: michael.welch at homepower.com
--^----------------------------------------------------------------
This email was sent to: michael.welch at homepower.com

EASY UNSUBSCRIBE click here: http://topica.com/u/?bz8Qcs.bz9JC9.bWljaGFl
Or send an email to: RE-wrenches-unsubscribe at topica.com

For Topica's complete suite of email marketing solutions visit:
http://www.topica.com/?p=TEXFOOTER
--^----------------------------------------------------------------